What patients ask
Patients ask the questions they would ask a friend, and they ask them in their own words: “good dentist in Tallinn for English speakers”, “how much does a dental implant cost in Estonia”, “which clinic does laser eye surgery and takes EU insurance”, “is X clinic any good”. These questions are local, commercial and often urgent, which makes them exactly the kind that retrieval-based engines answer with a shortlist and specifics. They are also the kind where a wrong answer costs a booking.
A clinic’s visibility check therefore needs questions in four groups: service questions (“who does X”), price questions (“how much does X cost”), local and practical questions (location, languages, insurance, new patients, hours), and brand questions about your clinic by name.
Accuracy comes first
For most businesses, being absent from an AI answer is the main problem. For a clinic, being misdescribed is worse. Engines state things confidently: an old price list, an address from before you moved, a treatment you stopped offering, a language your staff do not speak, or a founding date and city copied from a different clinic with a similar name. A patient who arrives expecting the quoted price, or who chose you for a service you do not provide, is a lost patient and a bad review.
There is also a regulatory dimension. Advertising rules for healthcare differ by country, and an engine that attributes a treatment outcome claim to your clinic can put words in your mouth that you would never publish. You cannot control what an engine writes, but you can control what the sources it reads say, and you can make sure your own site states the facts plainly enough to be the source that wins.
Run your brand questions first, on each engine, more than once. Write down every claim about your clinic. Mark each as correct, outdated or wrong, and note the source shown. That list is your first fix list.
Where clinic answers come from
When you look at the sources behind the answers (Perplexity shows them on every answer; ChatGPT and Claude show them when they retrieve), the same kinds of pages recur for clinics: map and business listings, health directories and booking platforms, review sites, professional registers, national health-service pages, and occasionally a clinic’s own site. Your own pages compete with these, and often lose, because a directory entry is structured and a clinic homepage is a slideshow.
Two consequences follow. Your listings on those platforms need to be complete, consistent and current, with the same name, address, services, languages and price ranges everywhere. And your site needs at least one plainly written page per service that says what it is, what it costs or the range, who it is for, where it is done and how to book.
The technical check for a clinic site
Clinic websites are often built on templates and booking plugins that render most of the content with JavaScript, and some are configured to block anything with “AI” in its user agent. Both make the clinic invisible to the engines that feed live answers. Check three things:
- robots.txt. The agents that feed answers are documented by OpenAI, Perplexity and Anthropic, and Google’s AI features use ordinary Googlebot crawling. Allow the search and user-triggered agents; decide separately about training crawlers.
- Text without JavaScript. Open your service pages with JavaScript disabled. If prices, services and the address disappear, the engines see a blank page.
- Structured data that matches reality. Organization or medical-organization markup with your legal name, address, phone and opening hours, consistent with your listings. Do not mark up claims you cannot stand behind.
Questions worth answering on your site
For each service, one page that opens by answering the obvious questions: what it costs (or the range and what determines it), how long it takes, who performs it, whether a consultation is needed first, which languages are spoken, and how to book. An FAQ block with these questions in the patient’s words helps the engine find the answer in your text. The tone should be the one you would use on the phone: direct, specific, without superlatives.
For local questions, a page per location with address, transport, parking, hours and the services available at that branch. Multi-location clinics are often recommended by one branch only because the other branches have no page an engine can read.
Measuring it
Because patients ask local and price questions, and because listings change, clinic visibility moves. Check the same questions the same way each month and look for three things: whether your clinic is named for the service and local questions that bring patients, whether what is said about you is correct, and whether your own pages are being cited or only third-party listings. Fix in that order.
Our free check reads your site, writes three questions a patient would ask, puts them to ChatGPT and Perplexity and emails you the answers with your clinic and competitors marked. The full audit runs fifteen questions twice on ChatGPT, Perplexity, Google AI Overviews and Claude, compares every claim with the facts you give us, checks the crawl gate and the citability of your service pages, and delivers a fix list a practice manager can hand to the web agency. The methodology describes exactly how the engines are queried, and the sample report is built around a dental clinic.